International Journal of Science and Research (IJSR)

ISSN (Online): 2319-7064

Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438

Model Reference Adaptive Control Based PID

Controller Designs for Steam Turbine Speed
Sherin A Kochummen1, Nasar A2
1
PG student, Department of Electrical & Electronics Engineering, TKM College of Engineering, Kollam, India

2
Assistant Professor, Department of Electrical & Electronics Engineering, TKM College of Engineering, Kollam, India

Abstract: A steam turbine, key part of the power plant that allows the conversion of the heat energy to electricity via mechanical
energy. Steam turbine control systems are being designed with todays technology to operate a turbine in a safe and reliable manner.
Tandem compound reheat steam turbine is introduced to increase the thermodynamic efficiency by multistage steam expansion.
Adaptation of PID controller using certain adaptation techniques improves the performance of the system. This paper investigates
Model Reference Adaptive Controller for speed control of Tandem compound reheat steam turbine using MIT Rule and Lyapunov rule
to bring up quick tracking and steady state control over the turbine speed thus comparing the result with conventional PID controller.

Keywords: Steam turbine, speed control, PID controller, Linear modeling of Steam Turbine, Model Reference Adaptive Control(MRAC),
MIT rule, Lyapunov rule.

1. Introduction for stability and good transient performance. There are

several prescriptive rules used in PID tuning. In this paper,
Steam turbines, one of the most versatile and oldest prime Model Reference Adaptive Control based PID controller is
mover technologies convert stored energy of high pressure designed for speed control of tandem compound reheat steam
and high temperature steam into rotary energy to drive a turbine based on MIT rule and Lyapunov rule and is
generator or mechanical machinery. Power generation using compared with with conventional PID controller response.
steam turbines has been in use for about hundred years, when
they replaced reciprocating steam engines due to their higher The structure of this paper is as follows. Section 2 describes
thermal efficiencies by multistage steam expansion and lower about speed governing in steam turbine. The linear
costs. The turbine may drive an electric generator or mathematical modelling of steam turbine is explained in section
equipment such as boiler feedwater pumps, process pumps, 3. Tuning of the conventional PID controller in section 4.
Section 5 discuses about Model Reference Adaptive
air compressors, paper mills and refrigeration chillers. The
Control(MRAC) and designing of PID controller for steam
thermodynamic cycle for the steam turbine, Rankine cycle is
turbine speed control system using MRAC based on MIT rule
the basis for conventional power generating stations where and Lyapunov rule. Simulation results is given in section 6.
water is first pumped to elevated pressure, which is medium Conclusion is discussed in section 7.
to high pressure depending on the type of turbine unit and
then most frequently superheated. The pressurized steam is
expanded to lower pressure in a multistage turbine, then
2. Speed Governing In Steam Turbine
exhausted either to a condenser at vacuum conditions or into
A typical governor model for steam turbines includes only
an intermediate temperature steam distribution system that
those components and control elements that are responsive to
delivers the steam to the industrial or commercial application
speed and speed reference and that supply an input signal to
and condensate is utilized back.
the control mechanism for the purpose of controlling speed
which are the control mechanism and steam control valve,
Various important milestones in the different types of steam
whose output is effective control valve area in response to
turbine modelling and speed control techniques were
speed deviation of the machine, and a section modeling the
proposed. In[3] the behavior of the shaft torque is simulated
turbine, whose input is steam flow and output is mechanical
with load and proportional control algorithm but with a
power applied to the generator rotor.
simplified first order system. Speed deviation control, with
Proportional(P) and Proportional Integral(PI) controllers, at
Capability of the speed control system to position the control
different load deviations and load set points is done in paper
valve so that a sustained oscillation of the turbine speed as
[10]. Adaptation of PID Controller using AI Technique like
produced by the speed control system, stays within a
fuzzy logic, Particle Swarm Optimization (PSO), Genetic
specified range during operation under steady-state load
Algorithm (GA), Bacterial Particle Swarm Optimization
demand or following a change to a new steady-state load
(BPSO) and Neuro fuzzy controller(ANFIS) for Speed
demand to protect the turbine from over speeding, monitors
Control of Isolated Steam Turbine was the work done by
all critical turbine parameters to avoid conditions that could
Mohamed .M. Ismail [4]. The turbine is considered just as a
cause equipment damage, to follow the load in a stable and
first order system along with an electro-hydraulic governor.
efficient manner.
PID controllers, the most commonly used controller
structures in industry present some challenges to control and
instrumentation engineers in the tuning of the gains required
Volume 4 Issue 8, August 2015
www.ijsr.net
Paper ID: SUB157255 319
Licensed Under Creative Commons Attribution CC BY
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438
The steam from this through the control valve enters the high steam [kg/m3], F is the steam mass flow rate [kg/s], t is time
pressure(HP) section of the turbine, develops some power in [sec.].
the high pressure section the steam temperature and pressure
thus reduced send to re heater after re heating the steam in Assuming the flow out of the turbine to be proportional to
the superheated condition again and then it enters the next pressure in the turbine
stage of the turbine is called medium pressure turbine (3)
section, MP. The steam leaving from the MP section enters
Where: P is pressure of steam in the turbine [kPa],
the low pressure (LP) sections through cross over pipings and
rated pressure; rated flow out of turbine.
then exhausted to the condenser as shown in figure 1.
With constant temperature in the turbine:
(4)

(5)

(6)

(7)

General model for a single steam turbine unit is

= (8)
Figure 1: Tandem compound steam turbine configuration
As the steam turbine configuration varies the complete
The turbine-generator coupling should run at a constant
turbine model can obtained accordingly. The complete model
speed for its safe operation, so we have to control the steam
transfer function can be obtained using the parameters from
flow to the turbine thus involves the use of control valves to
the following table.
modulate the steam flow by adjusting its position.
Table 1: Steam turbine parameters
3. Linear Modelling Of Steam Turbine
The steam, generated in the boiler flows through the various
stages of steam turbine and then ultimately exhausted to the
condenser. In order to develop the model a simple steam
The overall transfer function of tandem compound reheat
turbine unit with volume V is considered.
steam turbine is
The mathematical modelling of the steam turbine unit,
developed based on the continuity equation[1,3]: (9)

Servomotor used in speed control systems converts the

electrical signal into angular displacement thus used for
actuating purpose by governing the valve opening. The
suitability of the motor for a particular application depends
on the characteristics, purpose and operating conditions of
the system to control with stability, the speed of the turbine at
power output between zero and maximum, inclusive, when
the generator is operating isolated.

= (10)

Figure 2: A basic steam turbine unit = (11)

Steam mass flow rate,

Thus the complete proposed system is
(1)
(12)
(2)
4. Tuning of Conventional PID Controller
In the above equations, W is the weight of steam in turbine
[kg], V is the volume of turbine [m3], density of The ZieglerNichols closed loop method of tuning a PID
Volume 4 Issue 8, August 2015
www.ijsr.net
Paper ID: SUB157255 320
Licensed Under Creative Commons Attribution CC BY
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438
controller. The P (proportional) gain, is then increased (from (y) and reference model output ( ).
zero) until it reaches the ultimate gain , at which the output of
the control loop oscillates with a constant amplitude and the Tracking error,
oscillation period are used to set the P, I, and D gains. The (14)
bode plot is plotted for the open loop transfer function.
One possibility to adjust parameters in a way that loss
The critical frequency obtained ( ) is 3.46 rad/sec, function
magnitude of the open loop transfer function is 0.0186
(15)
Ultimate gain, = 92.047 To make J small, negative gradient make the update rule,
(16)
Oscillation period, = 1.8159
= controller parameter vector
The parameters of PID controller, = Adaptive gain
= = 54.145
=
=

Bode Diagram
Gm = 48.6 dB (at 3.46 rad/s) , Pm = 67.3 deg (at 0.0737 rad/s)
50

0
Magnitude (dB)

-50

-100

Figure 4: Basic block diagram of a MRAC system

-150
-90
If the process is linear with transfer function and K is
-135 unknown. The underlying design gives a system with transfer
Phase (deg)

function , where is a known parameter.

-180

-225 (17)

-270
-3 -2 -1 0 1 2
Defining a control law,
10 10 10 10 10 10
Frequency (rad/s)
(18)
Figure 3: Bode plot of steam turbine system
= = (19)
Transfer function of PID controller
= (20)
PID(s) = s)
= s) (13) 5.1 Design of PID Controller using MIT rule

When the parameter of any systems changes with respect to

5. Model Reference Adaptive Control time then the conventional controller action is not effective.
In case of MRAC based design the adjustable parameters
The Model Reference Adaptive Systems(MRAS) derived for corresponding to changes in plant will be determined by
deterministic continuous-time signals is an important referring to reference model specifying the property of
adaptive control where desired performance expressed in desired control system. For designing purpose the reduced
terms of reference model thus responding to the command order model of our plant is
signal, other than the normal feedback loop there is another
to change the controller parameters with respect to the error. (21)
The parameter adjustment mechanism can be gradient
method or by using stability theory[2].
And the reference model chosen here is
(22)
The MIT(Massachusetts institute of technology) rule central
to the adaptive nature of the controller aims to minimize the
squared model cost function by which the error function The characteristics or system parameters are usually not
minimized for perfect tracking between actual plant output always the same. For these reasons, tuning of the traditional
PID controller parameters to control this system for the
Volume 4 Issue 8, August 2015
www.ijsr.net
Paper ID: SUB157255 321
Licensed Under Creative Commons Attribution CC BY
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438
required performance faces a strong challenge, so it is better - (35)
to adapt the tuning parameters by an adaptation rule.

PID Controller output, - (36)

) (23) (37)

Plant output, = (38)

Y(s) = (24)
- (39)
Error =
Thus integrating these values we obtain adapted parameter
= (25) values which can direct plant output o desired output.

= (26)
= (27)
= (28)
= (29)
Approximating with reference model
=
6. Simulation Results
The simulation results were presented to illustrate the
performance, and to compare the proposed strategy with the
existing PID controller. Figures 8 provide the responses of
the PID controller under disturbances in load for different
operating conditions. Test results demonstrated the feasibility
and effectiveness of proposed approach. Its performance
comparison proved its superior disturbance rejection
response to variations in load and reference pressure. The
convergence rate depends directly on the value of adaptation
gain, . Simulation results indicate it is true for small values
of but behavior is quite unpredictable for large values and
so the selection of is crucial.
Speed Output using conventional PID controller
2

5.2 Design of PID Controller With Lyapunov Rule 1.5

Speed(rpm)

The adjustment rule obtained by Lyapunov rule is simpler.

The proportional, integral and derivative parameters in 1
control law are adapted using the adjustment mechanism. Model Output
0.5 Plant Output
Controller output,
(30)
0
0 50 100 150 200 250
Let plant transfer function is Time (seconds)
G(s) = ( (31) Figure 5: Speed Control using Conventional PID controller

= ( + - - y) (32) MRAC with PID controller using Lyapunov Rule

150

Let model transfer function, 100

Speed(rpm)

= (33)
50 Model Output
= 64 16 (34)
Plant Output
0
Error =

From the Lyapunov function, V(e, , , , , ) -50

0 2 4 6 8 10
Time (seconds)
Therefore,
Figure 6: Speed control using MRAC Lyapunov rule with
PID controller
Volume 4 Issue 8, August 2015
www.ijsr.net
Paper ID: SUB157255 322
Licensed Under Creative Commons Attribution CC BY
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438
MRAC with PID controller using MIT rule [4] Mohamed. M. Ismail, Adaptation of PID controller
150 using AI techniques for speed control of isolated steam
turbine, IEEE Journal of Control, Automation and
Systems, vol. 1, no. 1,2012.
Speed(rpm)

100 [5] RekhaRajan, MuhammedSalih. P, N. Anilkumar,

Speed Controller design for Steam Turbine,
International Journal of Advanced Research in
50 Model Output Electrical, Electronics and Instrumentation
Plant Output EngineeringJ., vol. 2, pp. 44004409, Sept. 2013.
[6] Sanjay Kr. Singh, D. Boolchandani, S.G. Modani,
0 NitishKatal, Multi-objective PID Optimization for
0 2 4 6 8 10 Speed Control of an Isolated Steam Turbineusing
Time (seconds) Gentic Algorithm, Research Journal of Applied
Figure 7: Speed control using MRAC MIT rule with PID Sciences, Engineering and Technology., vol. 7, pp.
controller 3441-3445, May. 2006.
SPEED OUTPUT WITH DISTURBANCE REJECTION [7] Ali Tarique, Hossam A. Gabbar, Particle Swarm
150 Optimization (PSO) Based Turbine Control,
Intelligent Control and Automation., vol. 4, pp. 126
137,May 2013.
SPEED(RPM)

100 [8] MirceaDulau ,DorinBica, Simulation of speed steam

turbine control system , International Conference
Inter disciplinarity in Engineering., vol. 12, pp. 716
Model Output 722, 2014.
50
Plant Output [9] Sumit Kumar Sar, Lillie Dewan, MRAC Based PI
Controller for Speed Control ofD.C. Motor Using Lab
View, wseas transactions on systems and control.,
0
0 2 4 6 8 10 vol.9 pp. 1015, 2014.
Time (seconds) [10] Beela Rajesh, and T. Padmavathi, Dynamic Analysis
Figure 8: Speed control using MRAC with PID controller of a fossil-fuelled steam electric power plant using
fuzzy PID controller,International Electrical
Engineering Journal (IEEJ), vol.3, no. 2, pp. 642-649,
Table 2: Performance comparison using different controllers 2012.
Adaptation mechanism Controller Peak Settling [11] Pankaj Swarnkar, Shailendra Jain, R. K. Nema Effect
Overshoot time(sec) of Adaptation Gain in Model Reference Adaptive
Controlled Second Order System, Engineering,
Zeigler-Nichols tuning PID 180 40 Technology & Applied Science Research,. vol. 1,pp.
MIT rule PID 103 4 70-75. 2011.
Lyapunov rule PID 102 4.5 [12] Swarnkar Pankaj, Jain Shailendra Kumar, Nema R.K
Comparative Analysis of MIT Rule and Lyapunov
7. Conclusion Rule in Model Reference Adaptive Control Scheme,
Innovative Systems Design and Engineering., vol. 2,
It is obvious that the MRAC controller,applied to more pp. 154162, 2011.
complicated problems which is designed based on the main [13] SuttipanLimanond, Kostas S. Tsakalis, Model
system, does not impose any overshoot to the system for an Reference Adaptive and Nonadaptive Control of
increase in reference pressure as the design requires Linear Time-Varying Plants, IEEE Transactions on
approximations to obtain sensitivity derivatives. The Automatic Control, vol. 45, no. 7, pp.1290-
convergence driving error to zero increases with increasing 1300,July2000.
gain. The turbine speed is achieved by coordinate control of
steam pressure with improved precision, reduced steady state Author Profile
error and fast response.
Sherin A Kochummen was born in Kerala, India in 26/10/1991.
References She received B.Tech degree in Electrical and Electronics
Engineering from Baselios Mathews II College of Engineering,
[1] Kundur P. Power system stability and control. Kollam, India in 2013.Currently, she is pursuing her M Tech degree
in Industrial Instrumentation & Control from TKM College of
McGraw-Hill Professional; 1994.
Engineering, Kollam, India. Her research interests include Control
[2] Astrom, K.J., and B. Wittenmark; Adaptive control; Systems. (sherinakoc@gmail.com)
2nd Edition: Prentice-Hall ,1994.
[3] MirceaDulau, DorinBica, Mathematical modelling
and simulation of the behaviour of the steam turbine,
International Conference Interdisciplinarity in
Engineering, pp. 723729, 2014.

Volume 4 Issue 8, August 2015

www.ijsr.net
Paper ID: SUB157255 323
Licensed Under Creative Commons Attribution CC BY